Alphabet earnings, agentic commerce and who leads AI visibility

0:02Hello everyone. Welcome to our eighth episode of Marketing to Agents. And with me as always, I have Lalit here. Hey Lalit.

0:11Hey Ashwani, good to be here.

0:13Hey, great to have you here. All right, that's perfect. So, this past week was of course as we all know, it was an earnings week. We had a lot of we had earnings from all the hyperscalers, all the max seven companies, most of them at least. And we heard some great updates coming out of that. So, why don't we kick off our today's episode with Alphabet's earning first since it will have direct implication on how and what agents are how and what the whole agent system is doing and what's happening on the AI search. So, would love to hear what are your thoughts on Alphabet's earnings and then yeah.

0:55Yeah, so from an AO point of view Ashwani, I think one thing that stood out in the Alphabet's earning call was that they they showed very high confidence and performance numbers on the AI mode and AI overview, right? So, you know, the AI mode has passed 1 billion monthly active users and you know, despite all the the you know, the whole chatter around the change in search behavior, their search revenue actually grew 17% year-over-year, right? So, it it now clearly you know, it's it's very clear that the AI overview and the AI mode are going to be a like a like a strong structural part of the Google's business, the Google search business. And there were also you know, like there there was also mention of that the the number of clicks, right? Like like let's Sundar Pichai clearly said that the AI searches are still sending billions of clicks to websites every week. Right, so so that to me was a like a standout argument which affects all of what we, you know, think and discuss about AO. That Google AI mode and AI overview are here to stay. They

2:12Are now driving significant business for Google, and publishers have to be now be comfortable with the reality where as, you know, where the Google zero term that we are all, you know, very familiar with where Google is not sending clicks. apparently, Google is sending clicks, but not sure which categories. So, this was for me the standout part from the Google earnings. What about you?

2:39Fantastic length, yeah, exactly. I think from a publisher point of view, this was a both a welcoming news and also a a question mark that if there is traffic going, then what kind of categories is the traffic going to? Because most of the time, at least we've been hearing from a number of brands across the board, lot of our users at Airpush have been reporting that there is definitely a shift in organic traffic. And it has not been completely compensated, of course, by the AI traffic which is coming from Google. So, definitely there is a a drop, but then how does that drop compensate with the increase coming from what you're exactly mentioned, what Sundar Pichai said? Now, the important point is if Google is showing a increase in search revenue, then does that mean it is because of increase in the number of clicks which is happening, or Google has categorically increased the costing or the price per click as well? So, I heard about something interesting, and I've still not completely validated it myself. So, as you would know, but for the longest period of time, for almost last 15 years since Google you know, PPC or paper click model had

3:59Launched, Google's click was always priced at the best second best price of a click. So, for example, for a click, if you were paying if you bid $5 for the click, and I'm the second bidder, and I bid only $2 for the click, then let it you if you won that click, you will only pay $2 and $0.01 for that click. You won't pay $5 for the click. That was Google's model for the last 15 years. But quietly early this year, Google has changed that model now, and now you will be paying $5 for that click. And I'll be paying if I win the click or I win the second click, I'll be paying $2 for that click. So, now So, some of these micro things have started changing about Google, which of course are not are not going to be talked about in the in the earnings call, but some of these things are happening. So, again, like I said, I have not really validated yet. We haven't been running huge budgets on on PPC, but we've been hearing this in the market that some of these changes are affecting. So, why when when Sundar Pichai is mentioning that there is increase in the search revenue, I'm not sure how much of that jump is coming

5:17Just from a click volume increase versus increase in the costing, increase in more inventory, adding more number of links. Organic for sure has gone significantly lower. So, again, mixed response. I think there's still more data which needs to be understood to be able to come to a conclusion on where this finally lands.

5:41Yeah, this this could be very well the the driver of the year-on-year year-on-year revenue growth. Because from all our usage patterns and what we, let's say, keep on discussing in the community, I think nobody's clicking. And which means, like, you know, a significant portion of people are not clicking. And and maybe whoever is clicking, they are probably costing the advertiser a lot more than earlier.

6:07Absolutely. Absolutely. And just one more quick data point around this fact is So, if you look at what change which Google made in May, which is now in Search Console, they're also showing impressions on AI searches. So, your brand impressions or your page citations on any AI overviews or AI searches which is happening on Google. So, that's a great way for Google to say, "Hey, your brand is still getting visibility." Or your links are still getting some visibility on Google. But, of course, that impression volume is minuscule as compared to the search volume impressions that we get today. And that's true for all the brands, right? So, the search volume impressions that any brand gets today is huge, and the AI impressions volume that is the difference is very, very significant there. Number one. And number two is now very interesting point, which is where these impressions of AI overviews. That's great that those impressions are there. But, when you go into your analytics and track, okay, how much traffic did those impressions generate or how much jump in brand queries you saw, that, of course, is a way, way, way narrower funnel now. Now, that So, a impression So, if an impression earlier at a on a good

7:25Position meant certain amount of clicks, that funnel has shrunk significantly, has narrowed significantly with AI overviews. So, yeah. Yeah, right. So, that's, I think, some of the things that we're learning now as Google is launching there Google is making like you said I think that we're going to be seeing more often now Google's AI mode and AI search is going to be combined and we'll see that lot more. but yeah, I think the traffic numbers will still be aggressively monitored.

8:00Yeah. But Ashwani like just from a user point of view do you think the click that happens in an AI answer is that actually worth more for the advertiser? Because from our from our let's say past discussions as well and from what we see from the community that the visitor that comes from the AI search is actually four times more likely to buy, right? They they are well researched, right? So so in in some sense I think if if a click happens there, right? It may be may be justified to charge more for that.

8:39100% I think that's a great point and there is data which is actually which is actually saying that is true that anyone who's who's already done their research on AI and then they're visiting your site there's a much higher propensity for them to complete the transaction. So you're right actually there it makes sense to charge higher for a user who's done their research and then coming to a platform. But yeah, does that Yeah.

9:07And you know what what we are seeing as a user behavior today is that people are spending a lot of time on the research in on the AI mode or AI mode in Google or any other chat engine so AI engine. And then they they would then either they would click something which was linked in the answer or they would open a new tab then you know search for a specific brand, right? Now the the whole behavior of let's say opening up a new tab is in itself frictionful. a lot of sort of money leakage from

11:32Advertiser point of view happening in the previous search era also, right? Where if you look at the eventual funnel conversion also, right? The the website click to conversion was always something that we were always focusing on improving. Then once you get the lead, then there's a good amount of effort on converting that lead to a a potential pipeline and then finally, you know, conversion, right? So, overall conversion for let's say if I take a general assumption across all categories, B2B, B2C you see, I think it was some roughly under 2 to 3% only, right? from let's say an ad click to a customer, right? So, so if if the AI experience is somewhere taking away all the user anxiety and information needs and landing you a high-intent customer, I think that click is worth a lot more. So, it may appear as a as a practice which probably is different from earlier, but I think is principally justified.

12:31So, lesser clicks, but higher revenue per click. That's how maybe that the math works for Google in that sense.

12:39Exactly. Yeah.

12:41Yeah. And just building on top of your Amazon anecdote, Amazon also reported their earnings today, and apparently their ads revenue is has jumped significantly as well. So, I think although although I think our hypothesis is that slowly discovery will move out of Amazon because if the discovery happens through LLMs, and then the final purchase happens on Amazon, then yes, but then I think still the dominant behavior today is still discovery also happens on Amazon. But yeah, that could be something we should change.

13:15Yeah. So, ad revenue also increases when the pool of advertisers increase, right? and for all these listing websites like, you know, marketplaces which list products, I think it is the volume of sellers is so so high there that, you know, you have like you have to be an advertiser to be to even to be even visible. So, these are sort of some natural trends which would which would anyway result in let's say increase in ad revenue. but but I would say these are all transient things. the eventual user experience, right? Which will which will be more stable user experience in the AI era would be where it will mostly transaction based. Yeah. And and you know, only if only few surface areas like Google, they would they would end up making, you know, revenue on advertisement. But all these marketplaces might still just make money on transaction.

14:12Agreed.

14:16Because if I'm an ad, right? If so if let's say because if I'm a brand, right? I I will have two choices. either I spend for my visibility on Amazon or I spend on my visibility on let's say ChatGPT or Google, right? And spending money on ChatGPT or Google would be much more profitable for the brands because they would they would they might actually land the visitor directly to their Shopify store, right? As opposed to an Amazon where even if the click happens, you're still landed in a like a a category page or you're on a product listing page, your competition is listed, right? So, as let's say overall commerce matures, I think Amazon is you know, like not well positioned to to sort of improve their e-commerce revenue. and that is why I think we discussed earlier also, right? That's why they're opening up their logistics arm, right? Their whole supply chain as a service. So, that if you look at how the e-commerce logistics and supply chain matured in India as a market because of the whole third-party revolution where third-party logistic players make it now super easy for brands to you know, just get get their ship products shipped to users without worrying too much about anything. I think the same doesn't exist in the

15:34US, right? So, the the world-class supply chain that Amazon has built is now they're now offering it as a API, right? To to individual sellers who have a Shopify store. That would be that that itself reflects like their own understanding of where the market is moving, right? So, for them to continue to play a significant role in the overall e-commerce economy, they they would have to bear that shift. And probably they would they would be big, you know, let's say they would gain a lot by being the preferred choice for sellers when it comes to let's say the logistics and supply chain.

16:14That's an interesting angle that if actually Yeah. So, another data coming out and that may be relevant to this discussion is that Etsy also reported their earnings and they specifically pointed out jump in their revenue and their transaction volume because of AI agentic commerce. So, apparently there is So, like you just took an example where there were actual listing pages Amazon listing pages in your in your response from the GPT. now imagine that if you were also okay with getting the agent to purchase the right product. Let's say you chose that and you just get the commerce done as well. So, I think we are seeing that some of those early signs also happening where agents are going ahead and and participating in the commerce as well. And could be interesting to see now how does the platform enable that or not? How does a platform like Etsy or Shopify or Amazon for that matter, how does that enable that layer of agentic commerce, too?

17:21Yeah, so over there I think the adoption of UCP, which is the universal commerce platform, I'm sorry, protocol, universal commerce protocol, I think that has seen significant adoption, right? And the UCP consortium is now much bigger and also includes other categories like travel, right? and I won't be surprised if you know, like tomorrow edtech also becomes part of it where where someone is, let's say, researching a course and then they end up, you know, buying, let's say, a a boot camp over, you know, on on the internet through an agent, right? So, so I think that is seeing good adoption and Shopify is leading the leading the charge there. Almost all Shopify stores, you know, come be the UCP is enabled by default. And Etsy, I you know, the thing about Etsy is that they they just happen to have made their catalog accessible to agents, right? They were probably late in the game, but then this basic optimization that led to a like roughly 6% increase, right? In you know, in traffic and and you know, discovery of products. And if you look at like a normal a consumer who enters the like the the world of the web, right? So, they

18:38Are not aware of all these brands, right? They they don't they're not aware whether an Etsy exists, they're not aware whether, let's say, another marketplace exists. They are aware about their life and their needs, right? And in the AI paradigm, this natural companion that we'll end up nurturing, which is the ChatGPT or any other AI assistant, right? They are extremely tireless, right? They can they can browse 2,000 websites and find what you what you need in your life, right? So, they As a user, I don't have to know whether so-and-so marketplace has exist or or right? So, I I just have to express my need to the to my AI assistant, and then they will surface the products that sort of fulfill my need, right? So, I would say it's a great let's say you know, era that we're entering, especially for individual sellers. It is disaggregates a lot of value, right? Which was captured in let's say aggregation platforms. and it would actually also help nurture a lot of niches, right? Where those niches don't exist because there's no demand to capture. There's no way to capture the demand, right? or they don't even know whether demand exists or not, right? So, niches would actually flourish a lot more, right? In this AI way of discovering products and services.

20:01I really like what I just said. moving on to our next item from my next item in the docket. So, there was a again, we're frequently now seeing studies coming out regarding AI search and what's happening on AI search. So, we had this new study come out this week. A couple of interesting findings from there. So, first of all, looks like and what this study covers is I think it's covers most likely e-commerce categories, but it says that for 85% of categories, there isn't a clear AI visibility leader right now. So, there aren't fixed brands which are getting listed on all different LLMs or AI searches. So, there So, that means those there are no clear winners still for 85% of categories. And the second interesting takeaway from that research or the report was that the brand search queries, so for example, for Airmeet or Air Plus, it's if there are brand search query volume that can be tracked on Google Analytics directly. If that is showing an increase, that is a great indicator for proxy for how good your AI search or AI visibility is.

21:20So, interestingly that that seems to have a great correlation with that. So, this again, we've been hearing AI searches increasing and certain categories are getting dominated by a few players. But, at the same time, there seems to be a large number of categories which are still open. And again, citation links, we know which are the usual websites which get quoted in citations on those links as well. So, a mixed bag of research findings from this report. any thoughts on this, Loven?

21:54Yeah, I think this was actually something of an eye-opener, right? Like at a at a very large scale, I think on the long tail of categories, peop- you know, what what you're getting in results in AI answers is still very probabilistic, right? So, the recurring nature of a brand being cited as a leader or as a top recommendation, that is not there, right? So, you pick any category, you search for you know, in your AI answer, like in some like there's a percentage of answers that would recommend, let's say, a a certain brand on as a top recommendation, right? And then another percentage would be where other brand is recommended as you know, as the top option. So, what this tells us is that that these are obviously probabilistic systems that we're dealing with. What it also tells us that the the mechanism with which they determine what is right in terms of their recommendation, right? What their recommendation logic per se, right? that obviously relies on their own training as well as data that is accessed during the the inference time, right? Yeah, so what I was saying is that it's it's largely probabilistic systems and the corpus of data which sort of determines or helps

23:10The the the AI model pick a pick a leader amongst the options, right? That corpus is also not pointed to a consensus, right? So, the it's an opportunity. It's an opportunity for brands to sort of gear switch gears in terms of publishing content which which which sort of validates their position as a market leader. And or whatever is their market position, right? it's a it's a great opportunity that the window is still wide open that bra- brands can double down on content creation in in obviously in a guided manner, like if if they're an AirPulse customer, AirPulse actually has a data-driven recommendation on which publishers are getting more preference and then, you know, work with those publishers to get the right content out. And also, AirPulse highlights the content gaps, which you know, from the point of view of your desired positioning, right? If let's say you're trying to optimize for certain prompts, which which which would only work if the AI systems understand you in a certain light. So, which means that there are some content gaps if you're not being cited or if you're not being, you know, mentioned, then AirPulse also highlights and AirPulse and a few other platforms also highlight the content

24:26Gaps. The good thing about AirPulse is that we we not just highlight the gaps, we also help you ship content to to fill the gaps. So, the space is still wide open for people to claim leadership. So, that's the key takeaway here.

24:42Exactly, and given that there are there are brands which have started investing in it, I this window, of course, is very time-bound window. So, I think taking an action on and again, going by what we've learned in the SEO era, it takes a while to get to get you to the top of the results, but at the same time with AI, that action that action might be sooner. But again, it still needs a lot of effort from each brand to get listed get mentioned on the AI answers. Perfect. And then just coming back to the second aspect where while there are brand mentions and while there is very hard way to know how many and in how many queries you're getting mentioned unless of course there are set of prompts that as a business as a brand I can track. But then what about everything else? Where am I getting mentions? What kind of pages am I getting mentions on? Now none of these none of the LLMs other than Google of course is sharing that data with us today. With a brand today. So one of the things which has come up in these studies is that a good proxy is to just track your brand search volume. So if you're seeing more and more increases in

26:02Your brand searches again either through your Google Search Console or through your analytics if you're seeing more organic traffic more traffic coming from your brand searches that could be a good proxy to see how your AI visibility is doing. But again, very rudimentary way very while there are great platforms which provide you very good visibility on specific prompts I think it's great to use platforms like Air PR to get that visibility clear and not leave that hanging or not leave that open for interpretation with a brand queries. Yeah.

26:41Yeah, so I think over there I think my my takeaway is that we were always looking for a ground truth metric. And because the visibility platforms especially the visibility monitoring platforms, they are basically looking at simulated user queries. And the the actual search queries data is not visible to us because the search surface area is also very fragmented today, right? people in my circle are not even going to ChatGPT. They're basically relying on their AI assistant, right? And which uses some search tool, which uses some index provider. So, extremely fragmented as we go forward. So, what is working or not working across all these engines is is hard to know for brands, right? So, the ground What this study reveals is that it probably is highlighting a usage pattern. The usage pattern is that when the AI ends up recommending a certain vendor option or a brand to the user, you know, at the end of their research time spent on AI, the natural next step for the user is to basically, you know, search for the brand website or do a branded query on the on the you know, on the next tab

28:01Or maybe go further down, right? In the in the AI queries in the AI session itself. And as we know that every user prompt is then split into fan out prompts, right? So, if a user is going deep into a certain brand's information, right? and by just asking more about a certain brand. For example, for [clears throat] Airmeet, right? Like if if a prospect is asking more about Airmeet, let's say Airmeet reviews, Airmeet versus Zoom, then the platforms would you know, the AI agents would invariably split that into fan out queries, right? And that would Those fan out queries would have a brand search you know, brand keyword in that, right? So, so that that was obviously like a like an implicit correlation, but now I think what is what has come out is that it it's there is a more more than a 50% correlation that if if you see increase in your brand search query volume in your own let's say Google search console or let's say your own search logs you can very confidently then assume that you know your AI visibility is also increasing on the side.

29:14Perfectly let yeah. I think you put it very accurately. So I think again both of these points do make a strong case for platforms like AirPulse to be able to help monitor, track, of course track your visibility AI visibility, understand where the gaps are and then of course help you act on those gaps as well. Prepare right kind of content, get your website in the right you know in the right structure technical technically ensuring everything is right to get the site be quoted more and your brand quoted more in AI searches. Perfect. Like I think that was it for our this week's episode. We did cover of course we will hear a lot more earnings calls this week which have happened which is Microsoft, Meta and again next week as well. But yeah, we'll keep bringing more and latest news back into this podcast. And thank you everyone for listening to this episode. Do share your feedback with us. And we'll see you again next week with another episode of Marketing to Agents. Thank you.

30:22Thanks Ashwani. Goodbye everyone.

30:24Bye.